Camelia, the Perl 6 bug

IRC log for #mojo, 2013-10-20

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:50 good_news_everyone joined #mojo
00:50 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/oAvDQw
00:50 good_news_everyone mojo/master df27d5a Sebastian Riedel: bring back some lost documentation
00:50 good_news_everyone left #mojo
01:21 dqw16 joined #mojo
01:22 good_news_everyone joined #mojo
01:22 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/C1i0Cw
01:22 good_news_everyone mojo/master 4f9c91e Sebastian Riedel: fixed small assignment bug in content helper
01:22 good_news_everyone left #mojo
01:25 good_news_everyone joined #mojo
01:25 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/cUYxHA
01:25 good_news_everyone mojo/master ea0a28f Sebastian Riedel: mention that named buffers are shared
01:25 good_news_everyone left #mojo
01:27 good_news_everyone joined #mojo
01:27 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/Dv7f9w
01:27 good_news_everyone mojo/master 1c394f3 Sebastian Riedel: no need to assign application
01:27 good_news_everyone left #mojo
01:53 bpmedley http://filebeagle.com <-- Woohoo! I love Mojolicious.
02:06 mixMonkey joined #mojo
02:36 ilbot2 joined #mojo
02:36 Topic for #mojo is now ūüé©, indubitably | http://mojolicio.us | http://irclog.perlgeek.de/mojo/today
02:37 Bender joined #mojo
02:43 BinGOs joined #mojo
02:43 garu joined #mojo
02:48 tba_ joined #mojo
02:48 perlite joined #mojo
02:48 mst joined #mojo
02:48 diegok joined #mojo
02:50 chansen_ joined #mojo
02:50 d4rkie joined #mojo
02:50 KindTwo joined #mojo
02:51 phillipadsmith joined #mojo
02:51 jzawodn_ joined #mojo
02:51 tianon joined #mojo
02:51 cosimo joined #mojo
02:51 jan joined #mojo
02:51 yko joined #mojo
02:51 og01 joined #mojo
02:52 phips joined #mojo
02:53 geira joined #mojo
02:54 Grrrr joined #mojo
02:55 mixMonkey2 joined #mojo
02:59 Grrrr joined #mojo
03:00 memowe joined #mojo
03:38 russum joined #mojo
03:40 preflex_ joined #mojo
03:58 dqw16 joined #mojo
04:10 russum1 joined #mojo
04:51 russum joined #mojo
05:08 dqw17 joined #mojo
05:08 russum1 joined #mojo
05:14 russum joined #mojo
05:30 russum1 joined #mojo
06:40 marcus bpmedley: Looks shiny
06:40 bpmedley Spanx
06:40 bpmedley Thanks, rather
06:41 * marcus runs and covers
06:42 marcus bpmedley: so, won't it be very expensive for you if people use this free offering for online backup?
06:42 bpmedley marcus: It's their HD.  I'm not supply the bandwidth or the hd space.
06:43 marcus bpmedley: oh, I see now
06:43 bpmedley How can I make that more clear?
06:43 KindTwo joined #mojo
06:44 marcus bpmedley: maybe not hide the explanation behind a '?' badge?
06:44 marcus 'Personal Cloud Storage' <- not obviously on your desktop
06:44 bpmedley Aha.  I'll have to think of another jingle.
06:45 bpmedley Thanks for the input!
06:47 marcus bpmedley: two more feedbacks: You should probably make a .app wrapper for osx
06:47 marcus bpmedley: and once you've started it and go to the web interface, you're kind of stuck
06:47 bpmedley Understood.  I'm still finalizing deployment.
06:47 bpmedley Sorry, stuck?
06:47 marcus bpmedley: There's a login dialog, and I don't have a u/p
06:47 marcus ?
06:48 bpmedley Did INSTALL.txt make sense?
06:48 marcus uh, didn't even see it
06:48 bpmedley I need to work on that point.  Will do.
06:49 marcus bpmedley: logging in as admin just says 'invalid credentials'.
06:49 marcus (with no password)
06:50 bpmedley Ahh.  Please try with a password.  This is awesome feedback.
06:51 marcus bpmedley: https://www.evernote.com/shard/s22/sh/6008‚Äč945f-1e30-4a21-a19b-dcfbbc9076af/d4be1e356‚Äčbe9fab34a65f04ac7dc3c8c/deep/0/Sparky.png # after trying admin / password
06:51 bpmedley Hrm.. is that with a password?
06:51 marcus bpmedley: tried password 'password'
06:52 bpmedley I believe the "i" gives details on what is needed for password strenght.
06:52 d4rkie joined #mojo
06:52 bpmedley That could be made clearer.
06:52 marcus lol
06:52 marcus that is totally incomprehensible
06:52 bpmedley heh
06:52 marcus a yellow badge with a message on it?
06:53 marcus that doesn't show before you click it...
06:53 bpmedley Sorry.  I'm learning.
06:53 marcus bpmedley: no worries.
06:53 marcus bpmedley: show error messages up front. in red
06:54 marcus managed to create a user now
06:54 bpmedley Cool
06:54 marcus oh, I didn't got to that next 'password must match screen :D
06:55 marcus there
06:55 bpmedley Sorry, I don't understand.  You never got the "Verify Password" prompt?
06:56 marcus bpmedley: I did. Thought I'd managed to add the user, but then understood I had to verify.
06:56 marcus bpmedley: I think just a 'register user' link on the frontpage with verify password in the registration dialog would be a lot clearer.
06:57 marcus bpmedley: once you're logged in it looks pretty neat tho :)
06:57 bpmedley Probably so.  I need to work on that.  Definitely something else.. :)
06:57 bpmedley Thank you.
06:57 marcus bpmedley: does it work through nats tho?
06:57 bpmedley No.. :(   Have to port forward.
07:30 Vandal joined #mojo
07:30 bpmedley marcus: http://68.180.67.102:8080/dashboard/shar‚Äčes/L1VzZXJzL2JwbS9EZXNrdG9wL2luaXQucG5n ¬†<-- Would that make things better if there was an "Initialization screen" when the db was created?
07:49 saurabh_abhishek joined #mojo
07:52 denis_boyun joined #mojo
08:07 mire_ joined #mojo
08:12 hrupp joined #mojo
08:25 KindTwo joined #mojo
08:26 gtodd joined #mojo
08:34 Eugene joined #mojo
08:36 Eugene Greetings. How can I access app in a model (that is used by controller)? Need to debug / log / dump something
08:37 Eugene model is a simple perl package
08:37 Eugene I guess I must add some "use .." but not sure which one
08:51 basiliscos joined #mojo
08:58 dotan joined #mojo
09:01 dod joined #mojo
09:01 sh4 joined #mojo
09:20 heytrav joined #mojo
09:20 dod joined #mojo
09:41 batman Eugene: you need to send the app object to the model object somehow.
09:42 batman but i must say that is a super bad design choice.
09:42 batman even if it's just for debugging
09:42 Eugene exactly, so I asked for better way
09:42 Eugene seems like a common task, no?
09:42 batman did you know that your controller has a "app" attribute?
09:42 batman you haven't described your common task
09:43 batman "something" is not very obvious
09:43 Eugene controller - yes, but not model and I want to be able to dump everywhere I need, otherwise debugging gets slow
09:43 Eugene something is some hash / array / scalar - anything
09:44 Eugene in my example it's a ua response
09:44 Eugene that is run in a model (that runs certain requests to another website)
09:46 batman MOJO_USERAGENT_DEBUG=1 morbo script/yourapp
09:46 Eugene I don't need source code, I need exact var dumped
09:48 * batman gives up
09:48 Eugene in particular I can see certain <div> in html, run dom->find which doesn't work, so I'm dumping this or that selector to see what's happening. Next time I will need to dump db result or mongo collection
09:48 Eugene pity
09:50 Eugene thanks anyway
09:50 Eugene you seem to be the only online :)
09:50 batman there's no magical dump whatever you want whenever you want it solution.
09:51 batman do you even know what the line i pasted does?
09:51 batman have you tried it?
09:51 batman you're also not telling us if this is production or development.
09:51 Eugene I have magic dump in my framework so that I can run $system->dump everywhere - so I want the same in Mojo :)
09:51 Eugene line you pasted - "MOJO_USERAGENT_DEBUG=1 morbo script/yourapp" ?
09:52 Eugene development
09:52 batman yes. that line
09:52 batman it will show you all the data between the ua and server.
09:53 Eugene yes, I use it from time to time but too much data. When you send 5-6 requests one after another and need to see exact collection made by certain css selector in certain response - it's not what you need
09:54 batman don't tell me what i need. you need to tell us what you need, and i'm giving up, since you're not doing that.
09:55 Eugene you're strange man :) I said I needed to be able to dump variables in a model
09:55 Eugene now I do it by passing app to a model which is not a good way as you noticed
09:55 Eugene noted
10:00 batman i have no idea why you're doing that.
10:01 Eugene I wish I saw how you debug your models then.. writing to temporary files?
10:04 batman http://blog.mongodb.org/post/59757‚Äč486344/faceted-search-with-mongodb # any idea why it's better to have the extra "facets1" sub doc, instead of just searching the actual values?
10:05 batman Eugene: i do stuff like this: https://github.com/jhthorsen/mandel/bl‚Äčob/master/lib/Mandel/Document.pm#L290
10:06 batman at the same time i write unit tests. i only DEBUG=1 when i don't understand why my tests fail.
10:07 Eugene yeah, but still sometimes you use it and seeing dump in browser is much better than having it in a file or console you have to switch to every time
10:07 denis_boyun joined #mojo
10:08 batman don't say "you". seriously, that's how i do it.
10:09 batman https://metacpan.org/module/Moj‚Äčolicious::Plugin::ChromeLogger ?
10:09 batman https://metacpan.org/module/Moj‚Äčolicious::Plugin::ConsoleLogger
10:10 Eugene doesn't make sense, I have no app to call
10:10 Eugene in a model
10:24 batman you have put yourself in a corner. escape from it, and you will find your solution.
10:24 batman i've been there many times myself. step back and draw a drawing of what you want.
10:25 Eugene writing tests now where I can use "say" :)
10:26 Eugene btw just noticed here - http://mojolicio.us/perldoc/Mojolici‚Äčous/Guides/Growing#Simplified_tests - they missed "use lib 'lib';" - otherwise it dies with "ouldn't find application class" error
10:27 jpn joined #mojo
10:28 batman no, they didn't
10:28 batman use prove
10:28 batman prove -l t/mytest.t
10:28 batman or just "prove -l" to run all tests
10:28 batman prove -vl if you want more diagnostics
10:28 Eugene ah, cool cheers
10:28 batman no problem.
10:35 * sri yawns
10:36 denis_boyun__ joined #mojo
10:45 basiliscos joined #mojo
10:49 basic6 joined #mojo
10:56 batman http://www.severalnines.com/blog/nosql-battle-ea‚Äčst-coast-benchmarking-mongodb-vs-tokumx-cluster
10:59 KindTwo joined #mojo
11:04 sri they so need to merge mongodb and tokumx
11:05 batman :)
11:05 batman did you see my question about facets?
11:06 sri don't understand the question
11:09 batman in the example, there a doc with {"language":"English", ...} but then they add {...,"facets1":{"name":"lan‚Äčguage","value":"English"}}
11:10 batman i don't understand why that's better than just searching the "language" key directly...
11:10 batman oops! "facets1":[ {...}, {...} ]
11:11 batman (array of hashes, not just a hash)
11:13 marcus bpmedley: yes, nicer.
11:13 dqw17 joined #mojo
11:20 stephan48 joined #mojo
11:31 hummeleBop joined #mojo
11:41 * sri still would like to see more code abstracted out of Mojo::UserAgent
11:51 batman is adam stokes here?
11:55 * sri wonders if the key to more abstraction would be more low level hooks
12:11 Eugene joined #mojo
12:11 sri hmmm
12:15 sri http/2 is still far away, maybe we should have a nicer hook to support socks in the meantime
12:16 sri something like $ua->on(connect => sub {...}) where you can assign a IO::Socket::Socks object to the transaction yourself whenever a new connection is needed (no keep-alive available)
12:21 dqw18 joined #mojo
12:26 sri a better solution might be to rip out all connection management
12:26 sri $ua->pool(Mojo::UserAgent::Pool->new)
12:31 sri nope, i have no idea how to do it right :(
12:38 basiliscos joined #mojo
12:38 hummeleBop joined #mojo
12:39 bowtie_ joined #mojo
13:05 dqw18 joined #mojo
13:07 denisboyun joined #mojo
13:09 plu joined #mojo
13:15 asarch joined #mojo
14:06 sh4 joined #mojo
14:11 mire joined #mojo
14:15 sh4|2 joined #mojo
14:16 sh4|2 joined #mojo
14:31 jberger joined #mojo
14:37 gryphon joined #mojo
14:47 denisboyun joined #mojo
14:59 denis_boyun joined #mojo
15:06 dqw18 joined #mojo
15:07 themage joined #mojo
15:07 themage hi people
15:08 themage question: any issues lately with Chrome and caching?
15:09 crab that's a strange question.
15:09 themage crab: I'm getting crazy today
15:09 themage so, I have this redirects on login/logout
15:10 jberger themage, getting timeouts and blank screens?
15:10 jberger I thought my netbook was flaking out on me
15:11 themage and chrome is asking for the pages I redirect to before the redirect - so I get 404s when I should be getting working pages and pages with logged in info where I should be getting anonymous pages
15:11 themage as far as I can tell Chrome is not asking for the redirected pages
15:11 themage and not respecting cache-control either
15:12 themage I found a couple of link online about this, but I'm not finding a proper workaround on this
15:14 themage (BTW, when I activate the browser inspector it starts behaving)
15:16 Eugene joined #mojo
15:42 tempire themage: chrome pre-fetches and caches pages
15:42 tempire it's one of the reasons it seems so fast
15:43 jberger themage, in your dev-tools settings, do you have "disable caching" checked?
15:45 jberger this is getting unusable!
15:45 jberger I'm glad to see I'm not alone
15:45 themage jberger: that would help me, but not solve the problem for my users
15:46 jberger no I agree
15:46 jberger I saw your comment that things improved with dev-tools open and I was wondering if that was why
15:46 themage and I don't have a problem with browsers caching pages, as long as they respect the cache-control pragmas
15:47 bpmedley There's a difference between caching and pre-fetching a page before a url is even entered.. :)
15:47 themage jberger: My guess (and httpry seems to confirm it) is that when you open the dev-tools it disables the pre-fetching
15:50 jberger makes sense
15:50 themage bpmedley: agree, but pre-fetching a page that is the result of the redirect of a post is weird, specially if your cookies change and your cache-control header says: max-age=0, no-cache, must-revalidate
15:51 bpmedley Sorry, I was trying to agree.  It's craziness.
15:51 themage (for both the redirect to page and the redirect it)
15:52 themage the only thing that seems to work is adding a random parameter that is never used for anything
15:53 jberger trying this out: http://www.technipages.com/goog‚Äčle-chrome-disable-prefetch.html
15:55 jberger seems to be helping for me
15:56 jberger bpmedley, congrats on the site btw
15:57 bpmedley Thanks! All feedback is welcome.
15:57 jberger I saw that you haven't solved nat yet, do that and the users will come
15:58 jberger that is how gotomeeting and other remote access products win over free solutions that don't do nat/redirect
15:59 bpmedley The only way I know to get around NAT is for the traffic to flow through a server I woudl control.  Is that correct?
16:00 basiliscos joined #mojo
16:03 bpmedley Other than port forwarding, of course.
16:11 good_news_everyone joined #mojo
16:11 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/B9YYdQ
16:11 good_news_everyone mojo/master a7f13df Sebastian Riedel: documentation tweaks
16:11 good_news_everyone left #mojo
16:12 sri hmmmmm
16:12 sri would be nice if there was a common api CONNECT support and a SOCKS addon could use
16:13 saurabh_abhishek joined #mojo
16:13 sri but how could that method possibly be turned into a public api? https://github.com/kraih/mojo/blob‚Äč/master/lib/Mojo/UserAgent.pm#L195
16:23 jberger bpmedley, yes, as far as I know
16:24 jberger maybe not
16:25 * jberger really doesn't know
16:25 bpmedley I do agree that the NAT issue needs to be solved to get a larger share of users.  Port forwarding won't work for a lot of people.
16:30 d4rkie joined #mojo
16:35 denis_boyun joined #mojo
16:49 sri i don't have high hopes... but have made a list post about SOCKS https://groups.google.com/forum/‚Äč#!topic/mojolicious/dlSgAOKQwE4
16:51 sri the curse of things mostly just working... people get less interested in low level hacking and the quality of contributions goes down drastically :/
16:51 sri or maybe the curse of "good enough"
16:53 sri oh well, at least something to point people at when they ask for SOCKS... i give up
16:55 d^_^b by SOCKS you mean like a SOCKS proxy support?
16:55 D4RK-PH0ENiX joined #mojo
17:03 jpn joined #mojo
17:04 sri what other SOCKS is there?
17:08 chansen sri: http://en.wikipedia.org/wi‚Äčki/File:Feet_in_socks.jpg
17:08 sri :D
17:11 d4rkie_ joined #mojo
17:17 d^_^b sri: no idea, after some reading looks that is what you were talking about.
17:18 d^_^b I don't have much experience in web, so didn't know if there was something new not related to SOCKS proxy.
17:35 russum joined #mojo
18:05 asarch joined #mojo
18:11 sh4 joined #mojo
18:37 * sri wonders what the best organized non-blocking http client might be
18:38 sri all the up and down grading can be really tedious
18:40 sri netty is not bad... but abstracted to the point where it's almost impossible to read without and ide
18:55 sri spray doesn't look too bad, but full of akka-isms
19:10 denisboyun joined #mojo
19:14 Mike-PerlRecruiter_ joined #mojo
19:39 good_news_everyone joined #mojo
19:39 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/4WxFLg
19:39 good_news_everyone mojo/master 8107e06 Sebastian Riedel: moved error handling from user agent to transaction
19:39 good_news_everyone left #mojo
19:52 denis_boyun joined #mojo
20:12 good_news_everyone joined #mojo
20:12 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/lZhtYQ
20:12 good_news_everyone mojo/master 492cdb7 Sebastian Riedel: document more boolean attribute values
20:12 good_news_everyone left #mojo
20:16 denisboyun joined #mojo
20:17 * sri is tempted to do a python project with so many async i/o goodies in 3.4 http://docs.python.org/dev/whatsnew/3.4.html
20:21 rem_lex|pivo joined #mojo
20:34 lukep joined #mojo
20:39 bpmedley Is embedding Mojolicious in another open source project allowed?
20:41 sri same as perl
20:42 marcus Just don't soil our good name ;)
20:42 marcus (What sri said)
20:42 bpmedley That's a very high standard to live up to.. :)
21:18 chansen sri: I think you may benefit from splitting part of the ::UA and ::Transaction code in to "transport" classes
21:19 sri chansen: exactly :) https://github.com/kraih/mojo/issues/423
21:22 chansen Well, why are you asking for input, when you know what needs to be done? ;P
21:33 sri chansen: a) i'm often wrong b) i have no idea how the api could look like exactly c) i'm wondering if there might be intermediate steps we could do and how proxy support will fit in
21:34 sri i've been looking through netty, spray and friends... but they are always way too complicated
21:36 sri well, spray is close, but goes crazy with actors
21:39 chansen mm, and netty is Java
21:46 sri i'd rather look at ruby or python code... but they are just not there yet ;p
21:47 basiliscos joined #mojo
21:49 chansen I would start by making the UA  as "dumb" as possible, UA should be responsible for finding the right transport and interchange messages (requests/responses), the underlaying transport should know how to send/receive messages
21:49 sri node.js is not organized enough, the spdy module is an extension of the https server... and so on
21:51 sri chansen: agree
21:54 bpmedley https://github.com/brianmed/sparky  <-- Here is the code for that filebeagle thing.
22:01 chansen what's "filebeagle thing"?
22:05 bpmedley http://filebeagle.com
22:08 chansen You probably need to expand the description, how would I benefit from your solution? How is it different from existing solutions?
22:08 bpmedley Agreed.  That's high on the priority list.
22:09 chansen I have no clue what your software does ;)
22:10 bpmedley Oh, sorry.  It's a little web server app that sits on your desktop and serves files and whatnot.
22:11 bpmedley It can take a directory of mp3's and make an m3u, pls, or use html5 audio to play them.
22:11 bpmedley Or just browse and download a file that is needed.
22:12 chansen Ah, ok
22:12 bpmedley There are a number of planned features.
22:14 chansen Cool, but where does the Cloud part come in?
22:15 bpmedley It's a Personal Cloud Storage app; like tonido: http://www.tonido.com.
22:15 bpmedley The goal is so you can access your files when away from home.
22:18 chansen Ah, you seriously need to update you site ;)
22:18 bpmedley The github or filebeagle.com?
22:19 chansen filebeagle.com if you want customers, github if you want collaborators =)
22:20 bpmedley What would you suggest for filebeagle.com?
22:22 chansen As a customer, I don't care if it runs on Windows, Linux or OS X, I assume it does until told otherwise. You need a punch line, what's unique with your software?
22:22 chansen Why should I download it?
22:22 bpmedley How about: Run your own cloud server so that your data is easily available on the go.
22:26 chansen How about something catchy "Your data, Where you want it!" or "Your data, Wherever you need it!"
22:26 bpmedley Hrmm.  I have catchy phrases right below it.  Would you reload the main page?
22:27 chansen Have you never worked with a marketing agency?
22:27 Ralesk bpmedley: hah, cloud server -- a buzzword, but people just don't know what it's about.  (besides, is it really the cloud if it's your PC?)
22:27 bpmedley No, I definitely need to.  This is my first main site and I got it going today.
22:28 bpmedley Ralesk: That's what zdnet was calling this type of software, actually.
22:28 sri "Like a cloud, but on your desktop!"
22:28 bpmedley We all need charmin.. :-0
22:28 Ralesk if there's a cloud on my desktop, I really hope it's not a cloud of smoke! :D
22:30 chansen bpmedley: Forget about the cloud, that's the secondary thing! The main thing is that I can access my data wherever I am! Forget about OS support, you cover all main OS, remove the logos and provide a punch line!
22:30 bpmedley Hrmm.  I'll need a better graphic.
22:31 bpmedley Perhaps a picture of a notebook, tablet, and smartphone.
22:31 bpmedley That's what I wanted - coudln't find a free one.
22:31 * sri still has no clue how the api between ua and transport could look like
22:32 chansen You shouldn't think about graphics, that's why we have programmers and Art Directors ;P
22:34 bpmedley Does channel sound odd to anyone else?
22:34 chansen sri: I would start as I said and parallely implement SPDY as it has more in common with HTTP than websockets
22:34 bpmedley I'm on a shoestring budget right now, tho.
22:35 sri chansen: i think factoring out connect proxy support might be a better first step
22:36 sri big refactoring has to wait for 5.0 anyway
22:37 * sri doesn't particularly care about spdy, and would rather implement the latest http/2 draft
22:38 sri what i really need is an idea for how socket and transport layers interact
22:39 chansen sri: the point isn't that it's SPDY, the point is that you need to think different inorder for the implementation to fit, and thus when HTTP/2.0 is stabilized, you will be more ready
22:39 sri sure
22:40 chansen and SPDY and HTTP has more in common than websockets
22:40 sri actually i wouldn't say that
22:40 sri SPDY is more like a layer below HTTP
22:40 sri while websocket is on the same layer
22:41 sri they have an odd relationship
22:45 chansen I would think otherwise, websockets receiving/sending messages using HTTP while SPDY is about sending requests/responses, subtle but important difference
22:46 sri the spdy protocol itself is also about receiving/sending async messages though
22:46 sri you have many different frame types
22:47 sri control frames and stuff
22:54 sri not that it reallt matters
22:57 asarch joined #mojo
23:00 chansen but from a higher POV (which is the key part from a Mojo POV) SPDY is closer to HTTP than websockets
23:01 sri it needs to be managed more like websockets though, due to reading *and* writing all the time ;p
23:03 sri i'm not any closer to having an api though :(
23:03 chansen Doesn't matter, the point is (from a client POV) I'm sending a HTTP request and receiving a HTTP response
23:04 sri and you multiplex requests over the same connection, that's something i've not thought about much yet either
23:05 chansen sri: forget about multiplex, you can't get that right until you have a design
23:06 chansen sometimes a design need's to be changed because of multiplex, but don't paint you self in a black corener
23:06 sri and then you get websockets over multiplexed http/2 connections \o/
23:06 * sri 's head explodes
23:07 chansen sri: you need to think outside your current box (Websockets and HTTP)
23:07 tudorconstantin joined #mojo
23:07 sri http/2 will be years of trial and error, that's for sure
23:08 sri chansen: i'm looking for small steps i could make now without api breakage to prepare for a transport layer
23:09 chansen fuck http/2, you are not listening
23:09 sri umm
23:10 sri chansen: what are you trying to say?
23:16 chansen What I tried to say is that SPDY is about sending/receiving requests/responses while Websockets is about sending messages using the HTTP transport, subtle but important difference.
23:25 TitanOfOld joined #mojo
23:26 * chansen give it some thoughts, I'll be after some needed sleep
23:33 dqw18 joined #mojo
23:40 batman joined #mojo

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary